Alibaba launches Token Foundry unit to unify AI model development
Synopsis
Alibaba has quietly created one of its most consequential AI organisational moves: a new unit called Token Foundry, placed directly under CEO Eddie Wu Yongming, merging its top model labs — a clear signal that AI commercialisation is now a C-suite priority, not a research side-project.
Key Takeaways
Alibaba Group Holding established Token Foundry on 8 June 2026 to consolidate AI model-development efforts.
Token Foundry merges Tongyi Lab and Future Life Lab and will be led directly by CEO Eddie Wu Yongming .
Future Life Lab developed Happy Horse (video-generation model) and Happy Oyster (open-world AI model) while under Taobao and Tmall Group .
Zhang Di , formerly of Kuaishou Technology 's Kling video-generation team, rejoined Alibaba in November 2025 to lead Future Life Lab .
Zhou Jingren will serve as Alibaba chief scientist and head the new Alibaba AI Future Research Institute , focused on frontier research.
Alibaba Group Holding on Monday, 8 June 2026 established a new division called Token Foundry, consolidating key model-development teams as the Chinese technology giant accelerates efforts to commercialise artificial intelligence and align its sprawling businesses around a single AI strategy.
What Token Foundry brings together
Token Foundry merges Alibaba's Tongyi Lab and Future Life Lab under one roof, according to people familiar with the matter. The new unit will be led directly by CEO Eddie Wu Yongming, signalling how central AI commercialisation has become to the company's top-level agenda. Future Life Lab was previously housed within Taobao and Tmall Group and developed projects including Happy Horse, a video-generation model, and Happy Oyster, an open-world AI model. The lab was overseen by Zheng Bo, chief scientist and vice-president of technology at Taobao and Tmall Group, who also serves as CTO of Alimama, Alibaba's digital marketing arm.Key personnel shaping the new structure
Zhang Di, who previously led development of Kuaishou Technology's video-generation model Kling, rejoined Alibaba in November 2025 to run Future Life Lab, reporting to Zheng Bo. His return underscores the competitive talent dynamics playing out across China's AI industry, where video-generation capabilities have emerged as a critical battleground. As part of the broader restructuring, Zhou Jingren will assume the role of Alibaba chief scientist and head a newly established Alibaba AI Future Research Institute. That institute will concentrate on frontier AI research and next-generation technological breakthroughs, creating a clear separation between applied commercialisation — Token Foundry's mandate — and long-horizon research.Why it matters
The move reflects a growing pattern among China's largest tech platforms: consolidating fragmented AI labs into tighter, CEO-proximate units capable of faster product cycles. By placing Token Foundry directly under Wu Yongming, Alibaba removes organisational layers that can slow the translation of model research into revenue-generating products.Competitive backdrop
Alibaba's restructuring arrives as domestic rivals including Baidu, ByteDance, and Tencent each pursue their own AI consolidation strategies. The video-generation segment, where both Happy Horse and Kling compete, has drawn particular attention as a near-term monetisation opportunity for foundation-model capabilities.What to watch next
Investors and industry observers will be watching whether Token Foundry's unified structure translates into faster product launches and measurable revenue contribution from Alibaba's AI portfolio. The parallel creation of the Alibaba AI Future Research Institute under Zhou Jingren will also be a signal of how aggressively the company intends to pursue foundational breakthroughs alongside near-term commercialisation.Point of View
Suggesting Chinese hyperscalers are converging on similar organisational logic. What mainstream coverage underplays is the talent signal: recruiting Zhang Di from Kuaishou's Kling team indicates that video generation is where Alibaba sees its nearest monetisation runway. The real test is whether CEO-level ownership of an AI unit accelerates revenue, or simply adds political weight to decisions that still move at institutional speed.
NationPress
24 Jul 2026
Frequently Asked Questions
What is Alibaba's Token Foundry?
Token Foundry is a new AI division established by Alibaba Group Holding on 8 June 2026 that consolidates Tongyi Lab and Future Life Lab under a single structure. It will be led directly by CEO Eddie Wu Yongming and is focused on commercialising Alibaba 's artificial intelligence capabilities.
Why did Alibaba create Token Foundry?
Alibaba created Token Foundry to unify fragmented AI model-development teams and accelerate the commercialisation of AI across its businesses. Placing the unit directly under CEO Eddie Wu Yongming signals that AI strategy is now a top-level corporate priority rather than a distributed research function.
Who leads Token Foundry and the new AI research institute?
CEO Eddie Wu Yongming will lead Token Foundry , while Zhou Jingren has been appointed Alibaba chief scientist to head the separate Alibaba AI Future Research Institute , which focuses on frontier research and next-generation breakthroughs.
What AI models did Future Life Lab develop?
Future Life Lab , previously part of Taobao and Tmall Group , developed Happy Horse , a video-generation model, and Happy Oyster , an open-world AI model. The lab was overseen by Zheng Bo and later run by Zhang Di , who rejoined Alibaba in November 2025 after leading Kuaishou Technology 's Kling video-generation model.
How does Token Foundry affect Alibaba's competition with Baidu and ByteDance?
Alibaba 's consolidation under Token Foundry is a direct response to the pace at which rivals like Baidu and ByteDance are shipping AI products. By removing organisational layers and centralising leadership under the CEO , Alibaba aims to shorten the cycle from model research to market-ready products, particularly in the competitive video-generation segment.
